sign in sign up
<<
Home , Australasian gannet, Northern gannet, Underwater videography

Vol. 442: 255–261, 2011 MARINE PROGRESS SERIES Published December 5 doi: 10.3354/meps09458 Mar Ecol Prog Ser

Dive strategies and foraging effort in the Australasian Morus serrator revealed by underwater videography

Gabriel E. Machovsky Capuska1,2,* , Robin L. Vaughn3, Bernd Würsig3, Gadi Katzir4,5, David Raubenheimer1,2

1Nutritional Ecology Research Group, Institute of Natural Sciences, Massey University, Private Bag 102 904 North Shore MSC, , 2Coastal-Marine Research Group, Institute of Natural Sciences, Massey University, Private Bag 102 904 North Shore MSC, Auckland, New Zealand 3Department of , Texas A&M University at Galveston, 200 Seawolf Parkway, Galveston, Texas 77553, USA 4Department of Marine Biology, University of Haifa, Mount Carmel, Haifa 31905, Israel 5Department of Evolutionary and Environmental Biology, University of Haifa, Mount Carmel, Haifa 31905, Israel

ABSTRACT: are specialist plunge divers that perform short and shallow V-shaped dives and long and deep U-shaped dives in pursuit of pelagic fish and . We used underwater videography to examine the patterns of behaviour and relative success rates of V- and U-shaped dives in Australasian gannets. A significantly greater proportion of U-shaped dives were associ- ated with successful prey capture than V-shaped dives (95% vs. 43%, respectively). The maxi- mum number of prey captured per dive by the gannets was higher than previously reported, reaching up to 5 fish in a single U-shaped dive. However, V-shaped dives were more efficient in terms of grams of prey captured per time spent underwater in successful dives. In contrast, a pop- ulation-level comparison of the mass of fish captured per total time spent under water (i.e. includ- ing unsuccessful dives) suggested that the 2 dive profiles were equally efficient. We also found that gannets adjusted their dive shape in relation to the depth of their prey rather than prey type, as previously hypothesized. Further studies are needed to understand decisions made by gannets while plunge diving in complex marine environments.

KEY WORDS: Gannets · Morus serrator · Dive shape · Prey capture success · Decision making · Multispecies feeding associations

Resale or republication not permitted without written consent of the publisher

INTRODUCTION 1866). Gannets plunge dive for fish and squid (Nel- son 1978), at times in feeding events called multi- Plunge diving has evolved as a highly specialized -feeding-associations (MSFA) that involve hunting technique among water families includ- other (e.g. shearwaters, gulls, terns), predatory ing gannets and (), tropicbirds (Phae - pelagic fish (e.g. tuna, ) and mammals (e.g. sea tho nidae), pelicans (Pelecanidae), gulls and terns lions, whales, dolphins). Some marine mammals herd (Laridae), and kingfishers (Cerylidae, Alcedinidae). fish towards the surface, where the fish remain To perform a plunge dive, an avian predator must within diving depth for (Camphuysen & first locate prey from the air and dive at high speeds Webb 1999). However, potential disadvantages to into the water for pursuit and capture (Cunningham plunge divers of these high-density associations

*Email: [email protected] © Inter-Research 2011 · www.int-res.com 256 Mar Ecol Prog Ser 442: 255–261, 2011

include competition (Clua & Grosvalet 2001), risk of Plunge diving in Northern gannets has been sug- predation (Heithaus & Frid 2003) and risk of acciden- gested to be a highly accurate foraging technique tal collision (Machovsky Capuska et al. 2011a). (Wanless et al. 2005), but no data on prey capture Until recently it was generally believed that gan- success are available for these species. nets hunt predominately by using ‘steep relatively Here we examine the association between dive vertical plunge diving from a considerable height’ profiles and hunting strategy in the Australasian (Nelson 1978, Garthe et al. 2000). However, by using gannet, using underwater videography. Our study motion data loggers it has been demonstrated that enabled us to test success rate (prey captured per Morus bassanus and dive) and efficiency (grams of fish captured per unit M. capensis use a variety of diving strategies includ- time underwater) of V- and U-shaped dives, and ing surface diving, plunge diving, and pursuit plung- examine the influence of marine mammals on dive ing (Ropert-Coudert et al. 2004, 2009). Northern and patterns. These analyses allowed a better under- Cape gannets display 2 dive types: (1) V-shaped standing of the conditions under which Australasian dives, which are shallow, of short duration, involve gannets adopt V- and U-shaped dive profiles. mostly the underwater momentum of the plunge, and may occasionally include a short phase of active propulsion by using wing flapping to pursue prey MATERIALS AND METHODS that had escaped the initial plunge, and (2) U-shaped dives, which are deeper and longer than the former The study was carried out from 24 August to dive type and always involve the bird shifting from 31 October 2005 and 8 to 12 August 2006 in Admi- the momentum phase to active propulsion by using ralty Bay (40° 57’ S, 173° 55’ E) and Basin wing flapping to pursue prey (Garthe et al. 2000, (40° 90’ S, 173° 90’ E), in the Marlborough Sounds, Ropert-Coudert et al. 2004, 2009). New Zealand. This region has been characterized by In Northern gannets it has been suggested that the high primary productivity due to unique local shape of the dive is related to type of prey, with V- oceanographic conditions (Heath 1985). These condi- shaped dives being used to capture larger pelagic tions underlie the presence of a large number of fish such as mackerel Scomber scombrus and herring marine mammals and seabirds (Markowitz et al. harengus with escape speeds of ca. 1.16 m s−1, 2004, Vaughn et al. 2008). The study was conducted and extended U-shaped dives being used for smaller under Texas A&M Use Protocol 2005-48. and slower pelagic fish such as capelin Mallotus villo- A total of 50 min of underwater video footage sus, with escape speeds of ca. 1.03 m s−1 (Garthe et al. of dive behaviour from 11 2000). However, the use of remote telemetry data log- stationary dusky dolphin Lagenorhynchus obscurus gers did not allow direct observations of the associa- feeding bouts were recorded. Video recordings were tion between prey type and hunting strategy (Garthe made using a combination of surface- swimming and et al. 2000), and consequently the question of why breath hold dives ranging in depths from 3 to 10 m gannets should employ V-shaped dives in some cir- (Vaughn et al. 2008), using a Sony DCR-HC 1000 cumstances and U-shaped dives in others remained (focal length 3.6 mm, shutter speed unresolved. In contrast, Elliott et al. (2008) suggested 1/500 s, 30 frames s–1) in an Amphi bico Invader elec- that the shape of a dive may instead be related to the tronic underwater housing (Vaughn et al. 2007). pursuit of prey schools at a specific depth. Footage was analyzed frame by frame using Adobe Australasian gannets Morus serrator are the second Premiere Pro CS4. Individual gannets were followed rarest member of Sulidae (Nelson 2005), and they are from the moment they penetrated the water to the closely related to the Northern and Cape gannets. moment they surfaced. The water surface and the They have been reported to dive up to 20.5 m and water bubble trajectory were used as vertical and hor- 23 m depths in New Zealand and Australian waters, izontal references. Dive depth was determined using respectively ( et al. 2009, Ismar 2010), although Adobe Photoshop CS4 Extended version 11.0.2. The they usually dive to about 2 m (Green et al. 2009). mean length of an adult Australasian gannet was Green et al. (2009) reported that Austra la sian gannets used as a size reference as it swam next to a prey ball remain submerged for a maximum of 42 s, although and perpendicular to the video camera. Nelson (1978) they routinely dive for less than 6 s. These gannets reported the mean of Australasian gannet body feed mainly on pilchard spp., En- length as 89 cm, but did not give the error around this graulis spp., saury Scomberesox spp., and jack mack- mean. Since the error is important for assessing the erel Trachurus spp. (Robertson 1992, Bunce 2001). accuracy of our method, we measured the length of Machovsky Capuska et al.: Dive strategies in Australasian gannets 257

20 dead Australasian gannets at autopsy, and ob- RESULTS tained a mean ± SD of 88.9 ± 5.0 cm. Given the close concordance of our measurement and that of Nelson From a total of 85 dives, mean dive duration was (1978), we used 89 cm as our reference value. 7.9 ± 5.4 s, and mean dive depth was 2.5 ± 2.0 m. Dives were categorized as V- or U-shaped (Fig. 1) More than 80% of dives were less than 4 m deep, and in accordance with Garthe et al. (2000). For each 80% of dives lasted less than 13 s (Fig. 2). U-shaped dive, the duration of the underwater momentum dives were longer (t-test, t = 13.758; df = 45; p < phase in which gannets descend through the water 0.0001) and deeper (t-test, t = 17.722; df = 45; p < column without wing propulsion, was compared with 0.0001) than V-shaped dives (Fig. 2). For V-shaped prey pursuit, in which gannets are propelled through dives (n= 39), the mean duration was 3.4 ± 1.8 s the water by actively moving their wings were com- (range = 1.1 ± 8.2 s) and mean depth was 2.9 ± 2.5 m pared. Dives were coded as successful if a fish was (range = 1.1 ± 4.0 m), whereas for U-shaped dives observed in the gannet’s . For both dive pat- (n = 46), the mean duration was 10.7 ± 5.3 s (range = terns, foraging efficiency (in g s–1 underwater) was 1.0 – 40.1 s), and mean depth was 4.0 ± 1. 5 m (range calculated, assuming a single prey item weighing = 1.0 − 9.1 m). While no significant differences were 32.5 g (Bunce 2001). We also calculated overall effi- observed in the average duration of the underwater ciency, for each , both in successful and momentum phase of V- and U-shaped dives (t-test, unsuccessful dives. p = 0.07), the wing flapping phase was longer in For statistical comparisons, data were tested using U-shape dives (t-test, t = –9.742, df = 65, p < 0.001) χ2, t-tests and 1-way analysis of variance (ANOVA), than in V-shaped dives. using PAWS Statistics version 18. We report data as We were unable to definitively identify the species means ± SD. of prey from the video footage, but the size and body shape was consistent with the small pelagic pilchard Sardinops neopilchardus. However, from a total of 11 feeding bouts studied, we identified 7 bouts (63%) in which the gannets continued to dive at fish schools as the schools changed depth within the water column.

Fig. 1. Morus serrator. The air bubble trajectories produced underwater by a diving gannet characteristic of: (a) V-shaped Fig. 2. Morus serrator. Frequency of U-shaped (n = 46) and dive and (b) U-shaped dive. Drawing not to scale. Illustration V-shaped dives (n = 39) of Australasian gannets relative to by Laura van Zonneveld (a) dive duration (seconds) and (b) depth (meters) 258 Mar Ecol Prog Ser 442: 255–261, 2011

In these cases the gannets used V-shaped dives (see also Green et al. 2009). Data on gannet diving when fish were at shallow depth and U-shaped dives strategies have previously been obtained by a variety when the fish were deep. This observation reveals a of techniques including devices attached to the birds, link between depth of prey and dive profile, while such as capillary tubes (Adams & Walter 1993), data controlling for prey type. loggers (Garthe et al. 2000, Hamer et al. 2000), and Prey capture success in U-shaped dives (95%) was from autonomous underwater vehicles (Brierley & higher than in V-shaped dives (43%) (χ2, F = 28.232; Fernandes 2001). While motion data loggers provide df = 1; p < 0.0001). Of the successful U- and V-shaped a sampling frequency high enough to represent an dives, respectively 7% and 5% included more than alternative to direct observations (Ropert-Coudert et one fish captured, with a maximum of 5 and 2 fish al. 2004), an efficient way to determine prey con- captured in a single dive. sumption rates in seabirds is to fit them with stomach A comparison of the efficiency of the 2 dive profiles loggers (Wilson et al. 1995). However, showed that the V-shaped profile was significantly the deployment of several devices on the body of a more efficient than the U-shaped profile (ANOVA, F flying is likely to interfere with natural = 16.628, df = 1, p < 0.0001) (Fig. 3). However, the behaviour (Phillips et al. 2003). Although underwater overall efficiency was similar for the 2 dive profiles. videography does not allow multiple comparisons of Since it was not possible to track an individual bird diving strategies of single individuals, it provides across successful and unsuccessful dives, we could a valuable high-resolution tool to explore at-sea only perform this calculation on a population-basis behavior in marine predators (Davis et al. 1999, (i.e. for each dive profile, we calculated g of fish cap- Takahashi et al. 2004, Grémillet et al. 2010), includ- tured in all dives divided by time spent underwater ing examination of patterns of diving in Australasian all dives). We were thus unable to calculate the error gannets highlighted in the present study. around this estimate.

Dive shape, foraging effort and prey capture success DISCUSSION Our findings that V-shaped dives were shallower Our analyses provide the first report of prey cap- and shorter in duration than U-shaped dives are con- ture success, as well as detailed evidence of dive sistent with the findings of Garthe et al. (2000). Max- strategies and foraging effort in Australasian gannets imum diving depths and durations recorded in the present study (9.13 m and 40.07 s) were less than those reported by Green et al. (2009; 23 m and 42 s) and Ismar (2010; 20.5 m) for Morus serrator. Dive depths recorded herein were also shallower than those found for M. capensis (12.5 m; Adams & Walter 1993) and M. bassanus (24 m; Ropert-Coudert et al. 2009). Our results also revealed that the momentum phase of the 2 dive profiles was on average simi- lar, and may be related to the biomechanics of the gannet’s body entering the water column (Ropert- Coudert et al. 2009). However, U-shaped dives were associated with long periods of active wing flapping pursuit and are thus likely energetically more ex - pensive than V-shaped dives (Ropert-Coudert et al. 2009). We acknowledge the possible limitation that our data were restricted by underwater visibility, whereas previous results were obtained using data Fig. 3. Morus serrator. Foraging efficiency in V- and U- loggers. shaped dives. The box plots show efficiency calculated as It has been suggested that dive profiles in seabirds mass of fish captured per unit time underwater during and may be used as an indicator of the successful dives (with the median, 25% and 75% quartiles, type of prey being pursued, but observational evi- error bars representing standard deviation). Bars sur- rounded by dashed lines show the mass of fish captured per dence for this is limited (Elliott et al. 2008). In the pre- unit of dive time both in successful and unsuccessful dives sent study, the resolution of the video footage did not Machovsky Capuska et al.: Dive strategies in Australasian gannets 259

allow a definitive identification of the prey species in Third, although gannets are visual predators (Lee & many instances, although the size and body shape Reddish 1981) that are able to see in the violet-sensi- were consistent with the small pelagic pilchard Sar - tive range of the spectrum (Machovsky Capuska et di nops neopilchardus. However, our results showed al. 2011b), it is still unclear how their visual mecha- that Australasian gannets altered their depth of for- nisms cope with reflection and refraction while aging to track the depth of a specific school of fish, detecting prey between the air and water interface. showing that in these circumstances dive profile is To date, studies on aerial and underwater visual acu- associated with foraging depth, as suggested by ity in pursuit diving birds involve several penguin Elliott et al. (2008). species (Sivak 1976, Sivak & Millodot 1977, Howland The relative efficiency of the 2 dive profiles re - & Sivak 1984) and also great Phalacroco- vealed a greater percentage of successful prey cap- rax carbo (Katzir & Howland 2003, Strod et al. 2008). tures and a greater maximum number of fish cap- Although gannets and cormorants are phylogeneti- tured during single U-shaped dives than has been cally related, their foraging strategies are rather dif- reported for any gannet species. On the other hand, ferent (Nelson 1978). Cormorants detect and pursue V-shaped dives were more efficient in terms of grams prey only after having submerged their head and of prey captured per second underwater during suc- eyes (Katzir & Howland 2003), whereas gannets cessful dives. However, the index of overall effi- detect their submerged prey from the air, plunge ciency per unit time underwater, including both suc- dive and then may switch to active pursuit in the cessful and unsuccessful dives, was similar for the 2 water column. dive profiles (Fig. 3). These results are, however, sub- ject to the caveat that we were unable to follow indi- vidual birds on both successful and unsuccessful Possible influence on dive profile by other dives and consequently our estimates of overall effi- marine predators ciency are calculated from population values and do not have associated error estimates. Our analysis of Australasian gannet dive types in Furthermore, there are other aspects relevant to relation to depth suggests that V- and U-shaped efficiency that remain to be quantified. First, in our dives could be indirectly related to the foraging be - comparisons we were unable to take into account the haviors of conspecifics and of other marine predators. time spent in the aerial phase of the dive. Second, Diving often takes place in multi-species feeding diving at shallow depths for a short period of time as sociations (MSFA) that involve a high density of may reduce the risk of gannets being captured by marine predators with different foraging tactics other predators (Crawford & Cooper 1996, Heithaus (Clua & Grosvalet 2001), and the species composition & Frid 2003) or the risk of injury or death due to acci- of these MSFAs might influence whether gannets use dental collisions (Machovsky Capuska et al. 2011a). V- or U-shaped dives. Thus, gannets might be more

Fig. 4. Morus serrator. A model of the use of V- and U-shaped dives related with different ecological variables. Statistically significant results are denoted with asterisks. Potential advantages and disadvantages are denoted by solid lines and dashed lines, respectively 260 Mar Ecol Prog Ser 442: 255–261, 2011

likely to use V-shaped dives when feeding with dol- eries. ICES J Mar Sci 58:904−915 phins, which herd the prey to shallower depths. Fur- Camphuysen CJ, Webb A (1999) Multi-species feeding associations in seabirds: jointly exploiting a ther, dolphin herding behavior at times appears to patchy environment. Ardea 87: 177−198 cause prey balls to become stationary, and it is likely Clua E, Grosvalet F (2001) Mixed-species feeding aggrega- easier for gannets to capture fish from stationary than tion of dolphins, large tunas and seabirds in the Azores. from mobile prey balls (Vaughn et al. 2008). Gannets Aquat Living Resour 14: 11−18 Crawford RJM, Cooper J (1996) Cape fur seal Arctocephalus might be more likely to use V-shaped dives when pusillus catches cape gannet Morus capensis ashore at feeding on stationary balls due to an improved ability Malgas Island. Mar Ornithol 24: 53−54 to detect and focus on the fish during a prey capture Cunningham RO (1866) On the Solan , or Gannet attempt. In contrast, when gannets aggregate in (Sula Bassana, Lim.). Ibis 2: 1−22 large numbers at a feeding bout, fish sometimes at- Davis RW, Fuiman LA, Williams TM, Collier SO and others (1999) Hunting behavior of a marine mammal beneath tempt to escape predation by descending to greater the Antarctic fast ice. Science 283: 993−996 depths (Vaughn et al. 2010). In these circumstances, Elliott KH, Woo K, Gaston AJ, Benvenuti S, Dall’Antonia L, gannets were observed in our study to shift from V- Davoren GK (2008) Seabird foraging behaviour indicates to U-shaped dives and continue their pursuit. prey type. Mar Ecol Prog Ser 354: 289−303 Garthe S, Benvenuti S, Montevecchi WA (2000) Pursuit Our study has provided several new insights into plunging by northern gannets (Sula bassana) feeding on the diving strategies of a plunge-diving predator, the capelin (Mallotus villosus). Proc R Soc Lond B 267: Australasian gannet (Fig. 4). The challenge ahead is 1717−1722 to integrate these with additional information and Green JA, White CR, Bunce A, Frappell PB, Butler PJ (2009) provide a model for gaining a broader understanding Energetic consequences of plunge diving in gannets. Endang Species Res 10:269−279 of the functional considerations underlying this inter- Grémillet D, Mullers RH, Moseley C, Pichegru L and others esting foraging strategy (Fig. 4). In particular, further (2010) Seabirds, fisheries, and cameras. Front Ecol Envi- studies are needed to understand the role of vision in ron 8: 401−402 the decisions made by gannets foraging in complex Hamer KC, Phillips RA, Wanless S, Harris MP, Wood AG (2000) Foraging ranges, diets and feeding locations of marine environments, energetic and nutritional con- gannets Morus bassanus in the North Sea: evidence from siderations, and the relationships between the for - satellite telemetry. Mar Ecol Prog Ser 200: 257−264 aging strategies of gannets and other marine preda- Heath RA (1985) A review of the physical of tors. Such questions demand an inter-disciplinary the seas around New Zealand-1982. NZ J Mar Freshw Res 19:79−124 ap proach that employs a variety of methodologies. Heithaus MR, Frid A (2003) Optimal diving under risk of predation. J Theor Biol 223: 79−92 Howland HC, Sivak JG (1984) Penguin vision in air and Acknowledgements. We thank D. and L. Boulton, N. water. Vision Res 24: 1905−1909 Duprey, H. and C. Pearson, D. Shelton, L. Timm, L. Watson, Ismar SMH (2010) Foraging and breeding ecology of the J. Weir, and numerous Earthwatch volunteers for assistance Australasian gannet Morus serrator, with applications for in the field, and Texas A&M University Department of rare New Zealand seabirds. PhD thesis, The University of Wildlife and Fisheries Sciences, Texas A&M University Auckland at Galveston’s Department of Marine Biology, the New Katzir G, Howland HC (2003) Corneal power and under- Zealand Department of Conservation, Marlborough District water accommodation in great cormorants (Phalacroco- Council, Earthwatch, National Geographic Society, and rax carbo sinensis). J Exp Biol 206: 833−841 E. Lee and L. Mooney travel grants for funding in the field. Lee DN, Reddish PE (1981) Plummeting gannets: a para- We also thank E. Libby, D. Rodríguez, E. Martínez and the digm of ecological optics. 293: 293−294 anonymous reviewers for helpful comments that greatly Machovsky Capuska GE, Dwyer SL, Alley MR, Stockin KA, improved the manuscript. G.M.C. is a recipient of the Insti- Raubenheimer D (2011a) Evidence for fatal collisions and tute of Natural Sciences (INS) at Massey University doctoral while plunge diving in Gannets. Ibis scholarship, and D.R. and G.M.C. are recipients of a grant 153: 631−635 from the Massey University Research Fund. Machovsky Capuska GE, Huynen L, Lambert D, Rauben- heimer D (2011b) UVS is rare in seabirds. Vision Res 51: 1333−1337 LITERATURE CITED Markowitz TM, Harlin AD, Würsig B, McFadden CJ (2004) Dusky dolphin foraging : overlap with aquacul- Adams NJ, Walter CB (1993) Maximum diving depths of ture in New Zealand. Aquat Conserv 14: 133−149 cape gannets. Condor 95: 734−736 Nelson JB (1978) The Sulidae: gannets and boobies. Oxford Brierley AS, Fernandes PG (2001) Diving depths of Northern University Press, Oxford gannets: acoustic observations of Sula bassana from and Nelson JB (2005) Pelicans, Cormorants and their relatives. autonomous underwater vehicle. Auk 118:529−534 Oxford University Press, Oxford Bunce A (2001) Prey consumption of Australasian gannets Phillips RA, Xavier JC, Croxall JP (2003) Effects of satellite (Morus serrator) breeding in Bay, southeast transmitters on and petrels. Auk 120: Australia, and potential overlap with commercial fish- 1082−1090 Machovsky Capuska et al.: Dive strategies in Australasian gannets 261

Robertson D (1992) Diet of the Australasian gannet Morus JP (2004) Penguin-mounted cameras glimpse underwa- serrator (G.R. Gray) around New Zealand. NZ J Ecol 16: ter group behaviour. Proc R Soc Lond B271: S281−S282 77−81 Vaughn RL, Shelton DE, Timm LL, Watson LA, Würsig B Ropert-Coudert Y, Grémillet D, Kato A, Ryan P, Naito Y, Le (2007) Dusky dolphin (Lagenorhynchus obscurus) feed- Maho Y (2004) A fine-scale time budget of Cape gannets ing tactics and multi-species associations. NZ J Mar provides insights into the foraging strategies of coastal Freshw Res 41:391−400 seabirds. Anim Behav 67: 985−992 Vaughn RL, Würsig B, Shelton DS, Timm LL, Watson LA Ropert-Coudert Y, Daunt F, Kato A, Ryan PG and others (2008) Dusky dolphins influence prey accessibility for (2009) Underwater wingbeats extend depth and duration seabirds in admiralty bay, New Zealand. J Mammal 89: of plunge dives in northern gannets Morus bassanus. 1051−1058 J Avian Biol 40: 380−387 Vaughn R, Würsig B, Packard J (2010) Dolphin prey herd- Sivak JG (1976) The role of the flat cornea in the amphibious ing: prey ball mobility relative to dolphin group and prey behaviour of the blackfoot penguin (Spheniscus demer- ball sizes, multispecies associates, and feeding duration. sus). Can J Zool 54:1341−1345 Mar Mamm Sci 26: 213−225 Sivak JG, Millodot M (1977) Optical performance of the Wanless S, Murray S, Harris MP (2005) The status of North- penguin eye in air and in water. J Comp Physiol A Neu- ern Gannet in Britain and Ireland in 2003/04. Br Birds 98: roethol Sens Neural Behav Physiol 119: 241−247 280−294 Strod T, Izhaki I, Arad Z, Katzir G (2008) Prey detection by Wilson RP, Pütz K, Grémillet D, Culik BM and others (1995) great (Phalacrocorax carbo sinensis) in clear Reliability of stomach temperature changes in determin- and in turbid water. J Exp Biol 211: 866−872 ing feeding characteristics of seabirds. J Exp Biol 198: Takahashi A, Sato K, Naito Y, Dunn M, Trathan PN, Croxall 1115−1135

Editorial responsibility: Hans Heinrich Janssen, Submitted: June 29, 2011; Accepted: October 21, 2011 Oldendorf/Luhe, Germany Proofs received from author(s): November 23, 2011